Databases

Reviewer: Robert M Curtice

As part of a special issue of Computer magazine to celebrate the 100th birthday of the IEEE, this article attempts to review the history and state-of-- the-art of databases in a mere 12 pages. The objective of such a whirlwind tour is unclear since anybody who knows the least bit about databases will find the treatment superficial to the point of annoyance, while anyone who knows nothing about the subject will probably be as confused after reading the article as before. Presumably, almost all readers of Computer fall into the first category. Perhaps the objective is to see how many technical terms can be mentioned in an article without defining them. All kidding aside, Professor Wiederhold has delivered a readable romp, given the limited space. Personally, I would have preferred less space devoted to the history of databases, since good historical summaries exist elsewhere, and the novice is not really interested in much detail about how we got where we are. As a consequence perhaps, I found some key ideas missing or treated much too lightly. End user facilities, application development aids, and data dictionaries are cases in point. It would seem almost a necessary consequence of such an article that many of the concepts presented are not quite accurate. Of course, the counter-argument might be that the flavor is essentially valid, the tradeoff between accuracy and getting the flavor across in the smallest space being made in favor of parsimony. Certainly the flavor of databases is presented. But the short detailed explanations allow far too much misinterpretation. Witness the following examples: :9BA Schema is used to retrieve data from the database. Schema information is compiled into a suite of programs that apply the basic append, read, and update operations to the data elements. Such an approach speeds up execution time. To eliminate the need for large numbers of programs, related elements may be grouped into segments. When performance is a major concern a database administrator can be used to control the quality and performance of the database. Other opportunities for misunderstanding abound. Several references are made to Cullinet and IDMS, but the impression is left that Total and Image are IBM products. Finally, we have this sentence following a discussion of the development of IMS on the IBM 360: “IBM gradually distributed IMS to 360 other users . . . .” I have the feeling this should read “other 360 users.” But I could be wrong.